- 3f97eb4 Introduce new UnarySDNode/BinarySDNode/TernarySDNode nodes, which coallocate by Chris Lattner · 19 years ago
- ab4ed59 eliminate the SDNode::setValueTypes method. by Chris Lattner · 19 years ago
- 63e3f14 eliminate a bunch of duplicate ctors and helper functions. by Chris Lattner · 19 years ago
- d429bcd move MorphNode to out of line and merge setNodeOperands into it. There is by Chris Lattner · 19 years ago
- 48b8592 simplify MorphNodeTo to take a VTList operand. by Chris Lattner · 19 years ago
- 67612a1 eliminate some extraneous methods in SDNode by Chris Lattner · 19 years ago
- c76e3c8 Give each selectiondag node class a home for it's vtable and rtti info by Chris Lattner · 19 years ago
- 9f24ad7 Switch VAlueMap from std::map to DenseMap. by Chris Lattner · 19 years ago
- 0da331f Switch NodeMap from std::map to DenseMap, this speeds up isel by 2.3% by Chris Lattner · 19 years ago
- 0ed4417 swtich vector-> smallvector, speeding up selectiondag stuff 1% by Chris Lattner · 19 years ago
- 40030bf Switch promoted/expanded ops over to using a DenseMap. Vector related maps by Chris Lattner · 19 years ago
- 718071c switch LegalizedNodes from std::map to a DenseMap. This speeds up isel by Chris Lattner · 19 years ago
- 00755df Eliminate some malloc traffic from LegalizeAllNodesNotLeadingTo, speeding by Chris Lattner · 19 years ago
- d48c5e8 Eliminate some std::sets. This speeds up isel of kimwitu by about 0.9% by Chris Lattner · 19 years ago
- cf5640b Switch SelectionDAG::ReplaceAllUsesOfValueWith to use a SmallSetVector for by Chris Lattner · 19 years ago
- 2ba528b switch the sched unit map over to use a DenseMap instead of std::map. This by Chris Lattner · 19 years ago
- 7971514 Switch ComputeTopDownOrdering over to using a densemap. This speeds up by Chris Lattner · 19 years ago
- 5efdecc Pasto by Evan Cheng · 19 years ago
- 832254e Changes to support making the shift instructions be true BinaryOperators. by Reid Spencer · 19 years ago
- b25fe82 Fixed uninitialized stuff inside LegalizeDAG. Fortunately, the only by Anton Korobeynikov · 19 years ago
- fea997a Fit in 80 columns by Chris Lattner · 19 years ago
- 527fae1 Emit a better assertion message for PR1133 by Chris Lattner · 19 years ago
- d385fd6 Allow the target to override the ISD::CondCode that's to be used to test the by Evan Cheng · 19 years ago
- 5cbf985 For PR1136: Rename GlobalVariable::isExternal as isDeclaration to avoid by Reid Spencer · 19 years ago
- 367f109 add initial support for handling inline asms with multiple constraints. by Chris Lattner · 19 years ago
- bcc5f36 Finish off bug 680, allowing targets to custom lower frame and return by Nate Begeman · 19 years ago
- 6aa279d Arguments are counting from 1. not from 0. Maybe we should change by Anton Korobeynikov · 19 years ago
- ac2b2cf More cleanup by Anton Korobeynikov · 19 years ago
- b10308e Propagate changes from my local tree. This patch includes: by Anton Korobeynikov · 19 years ago
- 44c3b9f Change the MachineDebugInfo to MachineModuleInfo to better reflect usage by Jim Laskey · 19 years ago
- 1ee2925 Make LABEL a builtin opcode. by Jim Laskey · 19 years ago
- de268f7 Renamed getTypeAlignmentShift() to getPreferredTypeAlignmentShift(). by Evan Cheng · 19 years ago
- f6d039a Remove the DoubleTy special case. by Evan Cheng · 19 years ago
- 24d6da5 For PR970: by Reid Spencer · 19 years ago
- 58092e3 Teach TargetData to handle 'preferred' alignment for each target, and use by Chris Lattner · 19 years ago
- 597a3bd Fix for PR1108: type of insert_vector_elt index operand is PtrVT, not MVT::i32. by Evan Cheng · 19 years ago
- c103057 For PR1043: by Reid Spencer · 19 years ago
- 42d7ccf Remove this xform: by Evan Cheng · 19 years ago
- c4e664b Fix PR1114 and CodeGen/Generic/2007-01-15-LoadSelectCycle.ll by being by Chris Lattner · 19 years ago
- 938ab02 add options to view the dags before the first or second pass of dag combine. by Chris Lattner · 19 years ago
- e5530da Compensate for loss of DerivedTypes.h in TargetLowering.h by Reid Spencer · 19 years ago
- 02114aa Move a function out of line. by Reid Spencer · 19 years ago
- 79cca50 Minor fix. by Evan Cheng · 19 years ago
- 5696622 Store default libgcc routine names and allow them to be redefined by target. by Evan Cheng · 19 years ago
- 6b6b6ef For PR1043: by Zhou Sheng · 19 years ago
- c8edc64 Naming consistency. by Evan Cheng · 19 years ago
- c62d4bb Fix for PR1075: bottom-up register-reduction scheduling actually increases register pressure. by Evan Cheng · 19 years ago
- ddae4bd Implement some trivial FP foldings when -enable-unsafe-fp-math is specified. by Chris Lattner · 19 years ago
- a4810b5 remove support for llvm.isunordered by Chris Lattner · 19 years ago
- 636c753 Expand fcopysign to the bitwise sequence if select is marked as expensive. by Evan Cheng · 19 years ago
- 068c5f4 Bug in ExpandFCOPYSIGNToBitwiseOps(). Clear the old sign bit of operand 0 by Evan Cheng · 19 years ago
- 489a87c CopyToReg source operand can be a register as well. e.g. Copy from GlobalBaseReg. by Evan Cheng · 19 years ago
- 0d630d2 GEP subscript is interpreted as a signed value. by Evan Cheng · 19 years ago
- f8e7a21 fix PowerPC/2007-01-04-ArgExtension.ll, a bug handling K&R prototypes with by Chris Lattner · 19 years ago
- 912095b Expand fcopysign to a series of bitwise of operations when it's profitable to by Evan Cheng · 19 years ago
- bcca340 Legalizer doesn't do an ANY_EXTEND if we don't ask for one so make sure by Reid Spencer · 19 years ago
- 376dd21 Restore previous behavior of defaulting to ZEXT. This works around two by Reid Spencer · 19 years ago
- 8c57dfb Two changes: by Reid Spencer · 19 years ago
- b47b25c Clean up from recent changes. Comment the new parameter to ExpandLibCall. by Reid Spencer · 19 years ago
- 4785781 For PR950: by Reid Spencer · 19 years ago
- e4d87aa For PR950: by Reid Spencer · 19 years ago
- 0b4f80e getLoad() and getStore() calls missed SVOffset operand. Thanks to Dan Gohman by Evan Cheng · 19 years ago
- cd3245a Eliminate static ctors from Statistics by Chris Lattner · 19 years ago
- 7df28dc May need to promote the operand (either sign_extend_inreg or and) before by Evan Cheng · 19 years ago
- 722cb36 LegalizeSetCCOperands() may end up inserting libcalls. They need to be by Evan Cheng · 19 years ago
- f2174da Fixed so that it dereferences the ostream pointer. by Bill Wendling · 19 years ago
- 5c7e326 Added an automatic cast to "std::ostream*" etc. from OStream. We then can by Bill Wendling · 19 years ago
- 36d4396 Fix PR1049 and CodeGen/Generic/2006-12-16-InlineAsmCrash.ll by Chris Lattner · 19 years ago
- e90460e Cannot combine an indexed load / store any further. by Evan Cheng · 19 years ago
- aa975c1 Expand FP undef by Evan Cheng · 19 years ago
- 0b1b9dc Allow promoted FP_TO_UINT / FP_TO_SINT to expand operand. by Evan Cheng · 19 years ago
- 966bf24 Expand fabs / fneg to and / xor. by Evan Cheng · 19 years ago
- 19103b1 Fix select_cc, select expansion to soft-fp bugs. by Evan Cheng · 19 years ago
- f6c4ccf This code was usurping the sextload expand in teh legalizer. Just make by Jim Laskey · 19 years ago
- 2d53a32 silence a bogus warning by Chris Lattner · 19 years ago
- 2b49c50 Expand FP compares to soft-fp call(s) by Evan Cheng · 19 years ago
- acd80ac 1. Tidy up jump table info. by Jim Laskey · 19 years ago
- 9f87788 More soft-fp work. by Evan Cheng · 19 years ago
- 548f611 Expand (f64 extload f32) to (f64 fp_ext (load f32)) if f64 type action is expand. by Evan Cheng · 19 years ago
- 98ff3b9 Expand fsqrt, fsin, and fcos to libcalls. by Evan Cheng · 19 years ago
- 6af00d5 Expand f32 / f64 to i32 / i64 conversion to soft-fp library calls. by Evan Cheng · 19 years ago
- 7b06bd5 Replace CastInst::createInferredCast calls with more accurate cast by Reid Spencer · 19 years ago
- 279101e Expand FP constant to integers if FP types are not legal. by Evan Cheng · 19 years ago
- 5c9ce18 Soft fp FNEG, SINT_TO_FP, UINT_TO_FP libcall expansion. by Evan Cheng · 19 years ago
- 0049521 Expand ConstantFP to load from CP if float types are being expanded. by Evan Cheng · 19 years ago
- 0ca6733 Expand i32/i64 CopyToReg f32/f64 to BIT_CONVERT + CopyToReg. by Evan Cheng · 19 years ago
- 7b2b5c8 - When expanding a bit_convert whose src operand is also to be expanded and by Evan Cheng · 19 years ago
- b15974a Expand formal arguments and call arguments recursively: e.g. f64 -> i64 -> 2 x i32. by Evan Cheng · 19 years ago
- f009483 fit in 80 cols by Chris Lattner · 19 years ago
- d9e06a5 this can only be fptrunc. by Chris Lattner · 19 years ago
- d93d46e Revert Nate's patch to fix X86/store-fp-constant.ll. With the dag combiner by Chris Lattner · 19 years ago
- 62be1a7 make this code more aggressive about turning store fpimm into store int imm. by Chris Lattner · 19 years ago
- 15f46d6 Change inferred cast creation calls to more specific cast creations. by Reid Spencer · 19 years ago
- 13acce3 Re-apply changes that were backed out and fix a naughty typo. by Evan Cheng · 19 years ago
- 2381f6f It seems the llvm::OStream class does not handle stream manipulators. by John Criswell · 19 years ago
- 70a24b3 Revert changes that broke oggenc on ppc by Chris Lattner · 19 years ago
- 25ece66 Don't convert store double C, Ptr to store long C, Ptr if i64 is not a legal type. by Evan Cheng · 19 years ago
- bbf1e5e f32 / f64 node is expanded to one i32 / i64 node. by Evan Cheng · 19 years ago
- b618230 Clean up some bad code. by Evan Cheng · 19 years ago
- 2cbba89 Move something that should be in the dag combiner from the legalizer to the by Nate Begeman · 19 years ago